Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberUS5991788 A
Publication typeGrant
Application numberUS 08/815,019
Publication dateNov 23, 1999
Filing dateMar 14, 1997
Priority dateMar 14, 1997
Fee statusPaid
Also published asUS6041340
Publication number08815019, 815019, US 5991788 A, US 5991788A, US-A-5991788, US5991788 A, US5991788A
InventorsLester Mintzer
Original AssigneeXilinx, Inc.
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
Method for configuring an FPGA for large FFTs and other vector rotation computations
US 5991788 A
Abstract
A method using replication of distributed arithmetic logic circuits and recursive interpolation of reduced angular increments of sine and cosine sum constants in logic look-up tables, permits the computation of vector rotation and large FFTs in a unitary field programmable gate array chip without required off-chip memory for storing constants.
Images(4)
Previous page
Next page
Claims(6)
I claim:
1. A method for configuring a field programmable gate array (FPGA) for carrying out a complex multiplication of the type (Xm -Xn)e-jΘ.sup.(k) where Xm and Xn are complex input variables and Θ(k)=2πk/N where N is a predetermined integer, the FPGA having an array of distributed arithmetic look-up-tables addressed by multiplier bits; the method comprising the following steps:
a) dividing said FPGA into at least four pipelined stages;
b) loading said look-up-tables with the sums of sin Θ and cos Θ factors logically ANDed with all possible combinations of said multiplier bits for π/2S increments of Θ in the angle range of 0 to π radians in a first stage of said FPGA, where S is an integer;
c) loading said look-up-tables with the sums of sin Θ and cos Θ factors logically ANDed with all possible combinations of said multiplier bits for π/22S increments of Θ in the angle range of 0 to π/2S radians in a second stage of said FPGA;
d) loading said look-up-tables with the sums of sin Θ and cos Θ factors logically ANDed with all possible combinations of said multiplier bits for π/23S increments of Θ in the angle range of 0 to π/22 S radians in a third stage of said FPGA;
e) loading said look-up-tables with the sums of sin Θ and cos Θ factors logically ANDED with all possible combinations of said multiplier bits for π/24S increments of Θ in the angle range of 0 to π/23S radians in a fourth stage of said FPGA;
f) applying said input variables to respective parallel-to-serial registers for said first stage;
g) inputting the difference of serial register stored input variables as address inputs to the look-up-tables for said first stage;
h) scaling and accumulating the outputs of the look-up-tables of said first stage and applying the outputs of said first stage as address inputs to the look-up-tables in said second stage;
i) scaling and accumulating the outputs of the look-up-tables of said second stage and applying the outputs of said second stage as address inputs to the look-up-tables in said third stage;
j) scaling and accumulating the outputs of the look-up-table tables of said third stage and applying the outputs of said third stage as address inputs to the look-up-tables in said fourth stage; and
k) storing real and imaginary trigonometric summed output of said look-up-tables of said stages in respective registers.
2. The method recited in claim 1 wherein said step g) comprises the steps of loading real components of said difference as addresses into a look-up-table and imaginary components of said difference as addresses into a look-up-table.
3. The method recited in claim 1 further comprising the steps of loading said look-up-tables with the sums of sin Θ and cos Θ factors logically ANDed with all possible combinations of said multiplier bits for π/25S increments of Θ in the angle range of 0 to π/24S radians in a fifth stage of FPGA.
4. A method of computing a rotation of a vector Xo Yo (where X and Y are complex) through a selected angle Θ using a digital device having distributed arithmetic table memory storing sums of trigonometric constant factors logically ANDed with combinations of multiplier bits; the method comprising the steps of:
a) determining the number of stages of vector rotation compatible with the capacity of said memory;
b) dividing said vector rotation into a plurality of successive angle rotation increments toward said selected angle Θ, said number of increments being equal to the number of said stages, each said increment being of successively smaller increments of said selected angle Θ of rotation;
c) carrying out computations in each said stage using corresponding memory-stored sums of trigonometric constant factors ANDed with said multiplier bits, said factors being selected for the corresponding increments of angle of rotation for the respective stage; and
d) scaling and accumulating the outputs of said table memory in each said stage and transferring the computational output of each such stage to the next successive stage as the input thereto and ultimately for the final stage as the computational output of said digital device.
5. A method for providing a gate-efficient implementation of the complex vector rotation defined by the term (X+jy)eJθ using a unitary field programmable gate array; the method comprising the following steps:
a) expressing the vector rotation in a sum of products form (X+jy)ejθ=xcos θ-ysin θ+j(ycos θ+xsin θ);
b) storing in distributed arithmetic look-up tables sums of sin Θ and cos Θ factors of said sum of products form for different values of θ;
c) employing successive angle rotation increments by partitioning angle ranges into successively smaller sets of increments of angle Θ to reduce the look-up table memory required in step b);
d) providing real and imaginary input variable components X and Y in bit serial form to successive stages of said look-up tables, each with said successively smaller angle increments;
e) using logic gates within said field programmable gate array to address said look-up tables.
6. The method recited in claim 5 wherein step e) is carried out by using the input variable components to address the first stage of said look-up tables and then using the output of each stage of look-up tables to address each successive stage of look-up tables, the output of the last stage being the parameters of the complex vector rotation.
Description
FIELD OF THE INVENTION

The present invention relates generally to the field of digital signal processing (DSP) in field programmable gate arrays (FPGAs) and more specifically to a method of carrying out complex vector rotation computations, such as large fast Fourier transforms (FFTs), in a single FPGA.

BACKGROUND OF THE INVENTION

The use of FPGAs for carrying out high speed DSP arithmetic computations has gained recognition in recent years. FPGAs having architectures which include logic blocks having multiple look-up-table function generators, such as the XC4000™ family of devices from XILINX, Inc., the assignee herein, are particularly suited for such computations. However, many of the important DSP algorithms are multiply-intensive, and even the FPGAs having the largest number of logic blocks, normally can't embed the multiplier circuits and the attendant control and support circuits in a single chip. It becomes incumbent on the designer to choose efficient DSP algorithms and to realize them with efficient circuit designs. The FFT is an outstanding example of an efficient DSP algorithm and distributed arithmetic is a well established design approach that replaces gate-consuming array multipliers with efficient shift and add equivalent circuits that offer comparable performance.

FFT--The discrete Fourier transform (DFT) of a sampled time series is closely related to the Fourier transform of the continuous waveform from which the time samples were taken. The DFT is thus particularly useful for digital power spectrum analysis and filtering. The FFT is a highly efficient procedure for computing the DFT of a time series and was reported by Cooley and Tukey in 1965 ("AN ALGORITHM FOR THE MACHINE CALCULATION OF COMPLEX FOURIER SERIES" by J. W. Cooley and J. W. Tukey, Math of Comput., Vol. 19, pp. 297-301, April 1965).

It takes advantage of the fact that the calculation of the coefficients of the DFT can be carried out interactively, which results in a considerable savings of computation time. If the time series contains N=2n samples, then for the N Fourier coefficients the FFT entails 2nN=2N log2 N multiply operations (assuming a radix 2 butterfly). In contrast, the DFT algorithm requires N2 multiply operations. The FFT advantage grows as N increases. Thus an 8 point DFT and FFT require 64 and 48 multiply operations, respectively, while an 8192 point DFT and FFT require 67.1106 and 212,384 multiply operations, respectively.

Distributed Arithmetic (DA)--Distributed Arithmetic was developed as an efficient computation scheme for digital signal processing (DSP). A United States patent describing this scheme is 1974 (U.S. Pat. No. 3,777,130 issued Dec. 3, 1974 entitled "DIGITAL FILTER FOR PCM ENCODED SIGNALS" by Croisier, D. J. Esteban, M. E. Levilion and V. Rizo. A comprehensive survey of DA applications in signal processing was made by White ("APPLICATIONS OF DISTRIBUTED ARITHMETIC TO DIGITAL SIGNAL PROCESSING: A TUTORIAL REVIEW" by S. A. White, IEEE ASSP Magazine, July 1989).

The distributed arithmetic computation algorithm is now being effectively applied to embed DSP functions in FPGAs, particularly those with coarse-grained look-up table architecture. Practical FIR, IIR and small size FFT designs have been developed. DA enables the replacement of the array multiplier, central to all these applications, with a gate-efficient serial/parallel multiplier with little or no reduction in speed.

DA makes extensive use of look-up tables (LUT's), thereby fully exploiting the LUT-based architecture of the Xilinx and other similarly structured FPGAs. The LUT used in a DA circuit will hereafter be called a DALUT. One can use a minimum set of DALUTs and adders in a sequential implementation to minimize cost. However, speed/cost tradeoffs can be made. Specifically, for higher speed, more DALUTs and adders may be employed. With enough DALUTs and adders, the range of tradeoffs extends to full parallel operation with all input bits applied simultaneously to the DALUTS and an output response generated at each system clock.

Distributed arithmetic differs from conventional arithmetic only in order in which it performs operations. The transition from conventional to distributed arithmetic is illustrated in FIGS. 1, 2 and 3. In FIG. 1 which illustrates conventional arithmetic, the sum of products equation, S=AK+BL+CM+DN is implemented with 4 serial/parallel multipliers operating concurrently to generate partial products. The full products are then summed in an adder tree to produce the final result, S. The functional blocks of the serial/parallel multiplier shown in the box of FIG. 1 include an array of 2-input AND gates with the A input derived from a parallel to serial shift register and the K input applied bit-parallel to all AND gates. A P bit parallel adder accepts the AND gate outputs addend inputs and passes the sum to an accumulator register. A divide by 2 block feeds back the register output to the augend inputs of the adder. In each clock cycle one bit of the serially organized data (Ai, Bi, Ci, Di) is ANDed with parallel operands (K, L, M, N) and four partial products are generated. Starting with the least significant serial bits, the partial products are stored in the four accumulator registers. On the next clock cycle, the next least significant bits again form partial products which are then added to the scaled by 1/2 previous partial product. The process repeats on successive clock cycles until the most significant bits have been shifted. When all the partial products, appropriately scaled, have been accumulated, they are fed to the adder array to produce the final output, S. Distributed arithmetic adds the partial products before, rather than after, scaling and accumulating them.

FIG. 2 shows the first embodiment of the distributed arithmetic technique. The number of shift and add circuits is reduced to one and is placed at the output of the array of simple adders, the number of simple adders remains the same. The two-input AND gates now precede the adders.

In a very important class of DSP applications known as linear, time-invariant systems the coefficients (K, L, M and N in our example) are constants. Consequently, the data presented to the shift-and-add circuit; namely, the output of the AND gates and the three simple adders depend only on the four shift register output bits. Replacing the AND gates and simple adders with a 16 word look-up table (DALUT) provides the final form (FIG. 3) of the distributed arithmetic implementation of the sum of products equation.

The DALUT contains the pre-computed values of all possible sums of coefficients weighted by the binary variables of the serial data (A, B, C and D) which previously constituted the second input to the AND gates. Now, with the four serial data sources serving as address lines to the DALUT, the DALUT contents may be tabulated as follows:

______________________________________A     B        C     D      address                              Content______________________________________0     0        0     0      0      00     0        0     1      1      N0     0        1     0      2      M0     0        1     1      3      M + N0     1        0     0      4      L0     1        0     1      5      L + N0     1        1     0      6      L + M0     1        1     1      7      L + M + N1     0        0     0      8      K1     0        0     1      9      K + N1     0        1     0      10     K + M1     0        1     1      11     K + M + N1     1        0     0      12     K + L1     1        0     1      13     K + L + N1     1        1     0      14     K + L + M1     1        1     1      15     K + L + M + N______________________________________

In general, the length or number of words in the DALUT is 2a where "a" is the number of address lines. The width, or number of bits per word cannot be precisely defined; it has an upper limit of b+log2a due to computation word growth where the coefficients are summed, as the content of the DALUT indicates (wherein b is the number of coefficient bits). The width of the table defines the coefficient accuracy and may not match the number of signal bits (e.g., the bits of A, B, C, and D) which define the dynamic range or linearity of the computation process.

Large FFTs in a Single FPGA--Now that the array multiplier has been replaced by a gate-efficient distributed circuit, there remains a second obstacle to overcome before a large size FFT can be practically embedded in a single FPGA, namely, the large memory required for the sine/cosine basis functions. An 8192 point FFT, for example, requires 8192 basis words and a look-up table capacity of 65,536 is needed for 16 bit accuracy. The Xilinx XC4000™ FPGA family with configurable logic blocks (CLBs) articulated as 321 look-up tables would require 131072/32 or 4096 CLBS. This number is almost double the CLB capacity (2,304) of the XC4062XL™ chip, currently the largest device in the XC400™ family.

It would be economically advantageous to provide a method for configuring smaller, less costly FPGAs such as the Xilinx XC4025™ device to perform 8192 or larger point FFTs and other vector rotations.

Those having skill in the relevant arts, will recognize that the invention hereinafter disclosed has some aspects which may be considered related to vector rotation iterations used in CORDIC processors, wherein CORDIC stands for Coordinate Rotation Digital Computer as disclosed in U.S. Pat. No. 5,371,753, assigned to Kuenemund et al. The CORDIC algorithm is similar to the invention disclosed in that both compute vector rotations 4096 words and cannot be reduced by any known interpolation scheme. A CORDIC FFT processor cannot be realized in a single FPGA.

SUMMARY OF THE INVENTION

This invention comprises the gate array implementation of the complex multiplication (x+jy)e-jθ where x and y are rectangular coordinates of a complex vector and θ is an angle of rotation. Since θ is expressed as increments of a basic, resolution angle, θo, it may be written as kθo, where k is an integer. In many applications, k takes on a multitude of values. In the case of the 8192 point FFT, k assumes 4092 values from 0 to 4091, and θo=2π/8192. There are accordingly, 4096 values of sine and cosine required for the complex multiplication which constitutes the preponderant FFT butterfly load. A single, large sine/cosine look-up table cannot be embedded in the largest FPGA. By decomposing θ (k) into sums of increasingly smaller angle segments, and by using DA computation techniques, the single large sine/cosine look-up table can be reduced to a set of small tables that can be embedded in a mid-size FPGA, such as the Xilinx XC4025™ device. The reduced DA look-up tables (DALUTS) of TABLE I bear this out. The complex multiply computations are iterative and are performed in pipelined DA stages that are nearly identical save for their look-up tables which reflect the angle segments.

                                  TABLE I__________________________________________________________________________DALUTS For Up to N + 8192 Point FFTXeR   Xel k  A1R        A1l            A2R                A2l                    A3R                        A3l                            A4R                                A4l__________________________________________________________________________0  0  0  0   0   0   0   0   0   0   00  1  0  0   1   0   1   0   1   0   11  0  0  1   0   1   0   1   0   1   01  1  0  1   1   1   1   1   1   1   10  0  1  0   0   0   0   0   0   0   00  1  1  0.3827        0.9239            0.0491                0.9988                    0.0061                        1   0.0008                                11  0  1  0.9239        -0.3827            0.9988                -0.0491                    1   -0.0061                            1   -0.00081  1  1  1.3066        0.5412            1.0479                0.9497                    1.0061                        0.9939                            1.0008                                0.99920  0  2  0   0   0   0   0   0   0   00  1  2  0.707        0.7070            0.0980                0.9952                    0.0123                        0.999                            0.0015                                11  0  2  0.707        -0.7070            0.9952                -0.0980                    0.9999                        -0.0123                            1   -0.00151  1  2  1.414        0   1.0932                0.8972                    1.0122                        0.9876                            1.0015                                0.99850  0  3  0   0   0   0   0   0   0   00  1  3  0.9239        0.3827            0.1467                0.9892                    0.0184                        0.9998                            0.0023                                11  0  3  0.3827        -0.9239            0.9892                -0.1467                    0.9998                        -0.0184                            1   -0.00231  1  3  1.3066        -0.5412            1.1359                0.8425                    1.0182                        0.9814                            1.0023                                0.99770  0  4  0   0   0   0   0   0   0   00  1  4  1   0   0.1951                0.9808                    0.0245                        0.9997                            0.0031                                11  0  4  0   -1  0.9808                -0.1951                    0.9997                        -0.0245                            1   -0.00311  1  4  1   -1  1.1759                0.7857                    1.0242                        0.9752                            1.0031                                0.99690  0  5  0   0   0   0   0   0   0   00  1  5  0.9239        -0.3827            0.2430                0.9700                    0.0307                        0.9995                            0.0038                                11  0  5  -0.3827        -0.9239            0.0700                -0.2430                    0.9995                        -0.0307                            1   -0.00381  1  5  0.5412        -1.3066            1.2130                0.7270                    1.0302                        0.9688                            1.0038                                0.99620  0  6  0   0   0   0   0   0   0   00  1  6  0.7070        -0.7070            0.2903                0.9569                    0.0368                        0.9993                            0.0046                                11  0  6  -0.7070        -0.7070            0.9569                -0.2903                    0.9993                        -0.0368                            1   -0.00461  1  6  0   -1.414            1.2472                0.6666                    1.0361                        0.9625                            1.0046                                0.99540  0  7  0   0   0   0   0   0   0   00  1  7  0.3827        -0.9239            0.3369                0.9415                    0.0429                        0.9991                            0.0054                                11  0  7  -0.9239        -0.3827            0.9415                -0.3369                    0.9991                        -0.0429                            1   -0.00541  1  7  -0.5412        -1.3066            1.2784                0.6046                    1.0420                        0.9562                            1.0054                                0.9946__________________________________________________________________________        .linevert split.            .linevert split.                    .linevert split.                            .linevert split.                                   .linevert split.DALUT    1R  .linevert split.  1l            .linevert split.  2R    2l                    .linevert split.  3R    3l                            .linevert split.  4R    4l                                   .linevert split.FFT Size:  16 ------>            .linevert split.                    .linevert split.                            .linevert split.                                   .linevert split.      128 ---------------------->                    .linevert split.                            .linevert split.                                   .linevert split.     1024 -------------------------------------->                            .linevert split.                                   .linevert split.     8192 ------------------------------------------------------>                                   .linevert split.
OBJECTS OF THE INVENTION

It is therefore a principal object of the present invention to provide a method for configuring a unitary FPGA to carry out large fast Fourier transforms without requiring off-chip sine/cosine memory.

It is another object of the invention to provide a method for configuring a unitary FPGA with a look-up-table to facilitate the computation of the complex products of vector rotation operations involving a large number of angle values between 0 and 2π radians.

It is still another object of the invention to provide a successive interpolation method for carrying out butterfly computations so that a sine/cosine look-up table and distributed arithmetic circuits could all be embedded in a unitary FPGA and produce large, fast Fourier transforms.

It is still another object of the invention to provide a successive interpolation method for carrying out complex multiplication computations so that a sine/cosine look-up table and distributed arithmetic circuits could all be embedded in a unitary FPGA and produce vector rotations.

BRIEF DESCRIPTION OF THE DRAWINGS

The aforementioned objects and advantages of the present invention as well as additional objects and advantages thereof will be more fully understood hereinafter as a result of the detailed description of a preferred embodiment when taken in conjunction with the following drawings in which:

FIG. 1 is a block diagram of a conventional four-product multiply and accumulate circuit;

FIG. 2 is a block diagram of the multiply and accumulate circuit of FIG. 1, but reorganized to reduce the number of shift and add operations;

FIG. 3 is a block diagram of a multiply and accumulate circuit in which a look-up-table replaces AND gates and adders of the circuit of FIG. 2;

FIG. 4 is an FFT Radix 2 Butterfly diagram or computation flow diagram referred to as a Butterfly computation because of the shape of the diagram;

FIG. 5 is a block diagram of the distributed arithmetic data path blocks of an FPGA for computing a Radix 2 Butterfly for a 16 point FFT; and

FIG. 6 is a block diagram of the distributed arithmetic data path blocks of a FPGA for computing a Radix 2 Butterfly for a 8,192 point FFT.

DETAILED DESCRIPTION OF THE DRAWINGS

The Radix 2 Butterfly

The invention may be understood by first referring to the radix 2 butterfly of FIG. 4. It should be noted, however, that for large transforms, higher order radices (e.g. 4 or 8) may offer greater efficiency--these, too, are amenable to DA implementation. The radix 2 butterfly computations can be expressed as follows:

Am =Xm +Xn =XRm +XRn +j[XIm +XIn ](Equation 1a)

An =(Xm -Xn)W k ; Wk =e-j2k/N N=2k                                                 (Equation 1b)

expanding Wk ;

where Xm and Xn are complex input variables and Am and An are the products of the butterfly configurations.

Wk =(e-j2π/N)k =COS Θk -j sin Θk ; Θk =2πk/N                                   (Equation 2)

By substituting Equation 2 into Equation 1b, a sum of products expression amenable to DA solution can be observed in Equation 3b below.

An =(XRm -XRn)COS.sup.Θk +(XIm -XIn) SIN.sup.Θk +j[(XRm -XRn)(-SIN.sup.Θk)+(XIm -XIn) COS.sup.Θk ]                                   (Equation 3)

Distributed Arithmetic

The sum-of-products (or vector dot product) equation wherein each term has a constant (e.g. coefficient) and variable (data) factors, defines the response of linear, time-invariant networks. Notable examples are FIR, and IIR filters and the discrete Fourier transform (DFT) and its efficient offspring the fast Fourier transform (FFT). Distributed arithmetic is a scheme for computing this sum-of-products without resorting to a full array multiplier. The derivation of the DA algorithm and a FIR filter design example is presented in an article entitled "A DISTRIBUTED ARITHMETIC APPROACH TO DESIGNING SCALABLE DSP CHIPS" by Bernie New in EDN magazine, page 107, Aug. 17, 1995. The multiplier is replaced by shift and add operations--a computation that is serial by multiplier bit (data) and parallel by multiplicand word (the constant factors). This technique is, of course, reminiscent of the software multiply technique of early computers. All of the serially multiplied product terms are computed concurrently; i.e., the partial product is the sum of the partial products of all the terms. The arithmetic sum of all partial products, namely, the sum of all the constant factors logically ANDed with their respective multiplier bits (1's and 0's) for all possible combinations of the multiplier bits is stored in a DALUT. A single DALUT serves all multiplier bits. Starting with the least significant data bits the DALUT is addressed and its contents stored in an accumulator register. The next significant bits address the same DALUT and its new contents are now added to one-half the previously stored value. The process repeats until the sign bits address the DALUT; its contents are subtracted from the stored sum yielding the response of the network. Thus, DA offers the gate efficiency of a serial implementation with performance approaching that of an array multiplier accumulator chip.

The dot product structure amenable to DA solution can be discerned in Equation 3. The x's are the input variables and the sinΘk, cos.sup.Θk represent the constant factors. If XR =(XRm -XRn) and XI =(XIm -XIn) and both serve as address inputs to the DALUT, then the DALUT for the real component of Equation 3b contains all possible combinations of XRb COS.sup.Θk +XIb SIN.sup.Θk while the DALUT for the imaginary component contains all possible combinations of XRb (-SIN.sup.Θk)+XIb COS.sup.Θk. Here the Xb's denote the multiplier bit values. Thus, for each value of k there are 4 words stored in each DALUT. Consequently, the 32 word CLB can provide the partial product contents for 8 values of Θ, i.e., k=0, 1 . . . , 7 --a 16 point FFT. The DA functional blocks for this 16 point FFT butterfly computer are shown in FIG. 5. The functional blocks consisting of serial adders and subtractors, parallel adders, shift registers, parallel-to-serial-registers, and truth tables (DALUTs) can all be readily configured in Xilinx CLBs. Excluding routing constraints, a total of 104 CLBs is needed. With a 50 MHZ system clock and 16 bit data, a butterfly is computed in 320 nsecs. And with N/2 log2N butterflies per FFT, the 16 point FFT can be computed in 10.24 usecs.

Large FFTs

The values of Θk cover the range of 0 to π in increments of 2π/N. Thus the 16 point FFT requires only 8 sin/cos values. These values are easily accommodated in a 32 word DALUT by allocating 3 address bits to Θk and 2 address bits to the real and imaginary data components. The Θk address bits cannot be scaled up without an exponential increase in DALUT size. However, by successive interpolation with increasingly smaller increments the DALUT is reduced to manageable size for even 8192 point FFTs. Let Θkk1k2k3 + . . .+Θkp with 8 values for each term; i.e., k=0-7 where Θk1 covers 0 to π in π/8 increments; Θk2 covers 0 to π/8 in π/64 increments; Θk3 covers 0 to π/64 in π/512 increments, etc.

The complex multiply, Equation 1b, now becomes: (((. . . ((Xm -Xn)e-jΘk1)e-jΘk2) . . . )e-jΘkp. This recursive expression can be implemented as a set of pipelined DA stages similar to the 16 point FFT (FIG. 5). A 16 bit data path (and sin/cos constants) and control functions can be realized with less than 100 CLBs per stage. An 8192 point FFT with 4096 values of Θk (12 address bits) requires only 12/3 or 4 stages--less than one half the CLB capacity of the Xilinx XC4025™ FPGA having 1024 CLBs. While one additional DA stage provides the capability to compute a 64K point FFT, the word growth due to computation gain and coefficient accuracy may set this as the practical upper limit for the Xilinx XC4025™ chip. However, there are presently several FPGAs having larger CLB capacity including the Xilinx XC4036EX™, XC4044EX™, XC4052XL™ and XC4062XL™ devices, as examples.

The data path functional blocks of the 4 stages are shown in FIG. 6. By using flip-flop registers resident in the CLBs, the combinational data flow can be limited to two CLB stages, thereby minimizing interstage timing uncertainties. The CLB count for each stage is indicated. With the total less than 400 CLBs, there are ample gate resources available for all system timing and data path controls as well as all off-chip data buffer address generation.

The basic computation for each successive interpolation stage of FIG. 6 is:

AR =XR COS Θk +XI SIN Θk 

AI =XI COS Θk -XR SIN Θk 

where the first stage is defined by Equation 3 wherein XR =XRm -XRn and XI =XIm -XIn. For all other stages, the x's are the outputs of preceding stages. With pipelining and with limited combination propagation delays, the 50 MHZ system clock can be retained with the same 320 nsec butterfly computation as the 16 point FFT. Thus, an 8192 point FFT can be computed in 17 msecs.

Distributed Arithmetic, when applied to replace array multipliers and to reduce trigonometric look-up tables by a novel successive interpolation scheme, affords a designer the opportunity to embed large FFTs in a single FPGA. Furthermore, these techniques may also be applied to the synthesis of precision sinusoids and to the design of complex mixers. It will now be understood that the present invention comprises a novel method for performing certain complex computations in FPGAs and other digital signal processing devices. A single FPGA configured for performing large (i.e., N=8,192) FFTs has been disclosed and is implemented by use of a series of successive interpolations with increasingly smaller angular increments for storing values and sums of trigonometric constants in look-up-tables. It will be also understood that the disclosed embodiments are of an exemplary nature and that the inventive method is conducive to even larger and more complex computations not explicitly disclosed herein. Accordingly, the present invention is to be limited only by the appended claims and their equivalents.

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US3777130 *Dec 15, 1971Dec 4, 1973IbmDigital filter for pcm encoded signals
US4680727 *Sep 24, 1984Jul 14, 1987Rockwell International CorporationComplex multiplier for binary two's complement numbers
US4970674 *Oct 3, 1989Nov 13, 1990Rockwell International CorporationProgrammable windowing FFT device with reduced memory requirements
US5033019 *Nov 21, 1989Jul 16, 1991Rockwell International CorporationVery-high-speed frequency-domain FFT windowing device
US5339265 *Aug 31, 1992Aug 16, 1994University Of Maryland At College ParkOptimal unified architectures for the real-time computation of time-recursive discrete sinusoidal transforms
US5371753 *Aug 26, 1993Dec 6, 1994Litton Systems, Inc.Laser diode mount
Non-Patent Citations
Reference
1Cooley, James W. and Tukey, John W., Apr. 1965, "An Algorithm for the Machine Calculation of Complex Fourier Series," Math of Comput., vol. 19, pp. 297-301.
2 *Cooley, James W. and Tukey, John W., Apr. 1965, An Algorithm for the Machine Calculation of Complex Fourier Series, Math of Comput., vol. 19, pp. 297 301.
3New, Bernie, Aug. 17, 1995, "A Distributed Arithmetic Approach to Designing Scalable DSP Chips," EDN, pp. 107-114.
4 *New, Bernie, Aug. 17, 1995, A Distributed Arithmetic Approach to Designing Scalable DSP Chips, EDN, pp. 107 114.
5White, Stanley A., Jul. 1989, "Applications of Distributed Arithmetic to Digital Signal Processing: A Tutorial Review," IEEE ASSP Magazine, pp. 4-19.
6 *White, Stanley A., Jul. 1989, Applications of Distributed Arithmetic to Digital Signal Processing: A Tutorial Review, IEEE ASSP Magazine, pp. 4 19.
Referenced by
Citing PatentFiling datePublication dateApplicantTitle
US6434582 *Jun 18, 1999Aug 13, 2002Advanced Micro Devices, Inc.Cosine algorithm for relatively small angles
US6499045Oct 21, 1999Dec 24, 2002Xilinx, Inc.Implementation of a two-dimensional wavelet transform
US6684235Nov 28, 2000Jan 27, 2004Xilinx, Inc.One-dimensional wavelet system and method
US6907439Mar 26, 2002Jun 14, 2005Lattice Semiconductor CorporationFFT address generation method and apparatus
US6976047Mar 28, 2002Dec 13, 2005Lattice Semiconductor CorporationSkipped carry incrementer for FFT address generation
US7119576Jun 18, 2004Oct 10, 2006Altera CorporationDevices and methods with programmable logic and digital signal processing regions
US7346644Aug 17, 2006Mar 18, 2008Altera CorporationDevices and methods with programmable logic and digital signal processing regions
US7430697 *Jul 21, 2005Sep 30, 2008Xilinx, Inc.Method of testing circuit blocks of a programmable logic device
US7536377 *Dec 18, 2003May 19, 2009Xilinx, Inc.Component naming
US7755641 *Aug 13, 2002Jul 13, 2010Broadcom CorporationMethod and system for decimating an indexed set of data elements
US7814137Jan 9, 2007Oct 12, 2010Altera CorporationCombined interpolation and decimation filter for programmable logic device
US7822799Jun 26, 2006Oct 26, 2010Altera CorporationAdder-rounder circuitry for specialized processing block in programmable logic device
US7836117Jul 18, 2006Nov 16, 2010Altera CorporationSpecialized processing block for programmable logic device
US7865541Jan 22, 2007Jan 4, 2011Altera CorporationConfiguring floating point operations in a programmable logic device
US7930336Dec 5, 2006Apr 19, 2011Altera CorporationLarge multiplier for programmable logic device
US7948267Feb 9, 2010May 24, 2011Altera CorporationEfficient rounding circuits and methods in configurable integrated circuit devices
US7949699Aug 30, 2007May 24, 2011Altera CorporationImplementation of decimation filter in integrated circuit device using ram-based data storage
US8041759Jun 5, 2006Oct 18, 2011Altera CorporationSpecialized processing block for programmable logic device
US8131793 *Nov 9, 2007Mar 6, 2012Pentomics, Inc.Efficient angle rotator configured for dynamic adjustment
US8244789Mar 14, 2008Aug 14, 2012Altera CorporationNormalization of floating point operations in a programmable integrated circuit device
US8255448Oct 2, 2008Aug 28, 2012Altera CorporationImplementing division in a programmable integrated circuit device
US8266198Jun 5, 2006Sep 11, 2012Altera CorporationSpecialized processing block for programmable logic device
US8266199Jun 5, 2006Sep 11, 2012Altera CorporationSpecialized processing block for programmable logic device
US8301681Jun 5, 2006Oct 30, 2012Altera CorporationSpecialized processing block for programmable logic device
US8307023Oct 10, 2008Nov 6, 2012Altera CorporationDSP block for implementing large multiplier on a programmable integrated circuit device
US8386550Sep 20, 2006Feb 26, 2013Altera CorporationMethod for configuring a finite impulse response filter in a programmable logic device
US8386553Mar 6, 2007Feb 26, 2013Altera CorporationLarge multiplier for programmable logic device
US8396914Sep 11, 2009Mar 12, 2013Altera CorporationMatrix decomposition in an integrated circuit device
US8412756Sep 11, 2009Apr 2, 2013Altera CorporationMulti-operand floating point operations in a programmable integrated circuit device
US8458243Mar 3, 2010Jun 4, 2013Altera CorporationDigital signal processing circuit blocks with support for systolic finite-impulse-response digital filtering
US8468192Mar 3, 2009Jun 18, 2013Altera CorporationImplementing multipliers in a programmable integrated circuit device
US8484265Mar 4, 2010Jul 9, 2013Altera CorporationAngular range reduction in an integrated circuit device
US8510354Mar 12, 2010Aug 13, 2013Altera CorporationCalculation of trigonometric functions in an integrated circuit device
US8539014Mar 25, 2010Sep 17, 2013Altera CorporationSolving linear matrices in an integrated circuit device
US8539016Feb 9, 2010Sep 17, 2013Altera CorporationQR decomposition in an integrated circuit device
US8543634Mar 30, 2012Sep 24, 2013Altera CorporationSpecialized processing block for programmable integrated circuit device
US8549055Mar 3, 2010Oct 1, 2013Altera CorporationModular digital signal processing circuitry with optionally usable, dedicated connections between modules of the circuitry
US8577951Aug 19, 2010Nov 5, 2013Altera CorporationMatrix operations in an integrated circuit device
US8589463Jun 25, 2010Nov 19, 2013Altera CorporationCalculation of trigonometric functions in an integrated circuit device
US8589465May 8, 2013Nov 19, 2013Altera CorporationDigital signal processing circuit blocks with support for systolic finite-impulse-response digital filtering
US8601044Mar 2, 2010Dec 3, 2013Altera CorporationDiscrete Fourier Transform in an integrated circuit device
US8620977Aug 7, 2013Dec 31, 2013Altera CorporationModular digital signal processing circuitry with optionally usable, dedicated connections between modules of the circuitry
US8620980Jan 26, 2010Dec 31, 2013Altera CorporationProgrammable device with specialized multiplier blocks
US8626815Mar 3, 2009Jan 7, 2014Altera CorporationConfiguring a programmable integrated circuit device to perform matrix multiplication
US8645449Mar 3, 2009Feb 4, 2014Altera CorporationCombined floating point adder and subtractor
US8645450Mar 2, 2007Feb 4, 2014Altera CorporationMultiplier-accumulator circuitry and methods
US8645451Mar 10, 2011Feb 4, 2014Altera CorporationDouble-clocked specialized processing block in an integrated circuit device
US8650231Nov 25, 2009Feb 11, 2014Altera CorporationConfiguring floating point operations in a programmable device
US8650236Aug 4, 2009Feb 11, 2014Altera CorporationHigh-rate interpolation or decimation filter in integrated circuit device
US8706790Mar 3, 2009Apr 22, 2014Altera CorporationImplementing mixed-precision floating-point operations in a programmable integrated circuit device
US8732225Oct 11, 2013May 20, 2014Altera CorporationDigital signal processing circuit blocks with support for systolic finite-impulse-response digital filtering
US8751551Nov 21, 2013Jun 10, 2014Altera CorporationModular digital signal processing circuitry with optionally usable, dedicated connections between modules of the circuitry
US8762443Nov 15, 2011Jun 24, 2014Altera CorporationMatrix operations in an integrated circuit device
US8788562Mar 8, 2011Jul 22, 2014Altera CorporationLarge multiplier for programmable logic device
US8805916Mar 3, 2009Aug 12, 2014Altera CorporationDigital signal processing circuitry with redundancy and bidirectional data paths
US8812573Jun 14, 2011Aug 19, 2014Altera CorporationCalculation of trigonometric functions in an integrated circuit device
US8812576Sep 12, 2011Aug 19, 2014Altera CorporationQR decomposition in an integrated circuit device
US8862650Nov 3, 2011Oct 14, 2014Altera CorporationCalculation of trigonometric functions in an integrated circuit device
US8886695Jul 10, 2012Nov 11, 2014Altera CorporationNormalization of floating point operations in a programmable integrated circuit device
US8886696Mar 3, 2009Nov 11, 2014Altera CorporationDigital signal processing circuitry with redundancy and ability to support larger multipliers
US8949298Sep 16, 2011Feb 3, 2015Altera CorporationComputing floating-point polynomials in an integrated circuit device
US8959137Nov 15, 2012Feb 17, 2015Altera CorporationImplementing large multipliers in a programmable integrated circuit device
US8996600Aug 3, 2012Mar 31, 2015Altera CorporationSpecialized processing block for implementing floating-point multiplier with subnormal operation support
US9053045Mar 8, 2013Jun 9, 2015Altera CorporationComputing floating-point polynomials in an integrated circuit device
US9063870Jan 17, 2013Jun 23, 2015Altera CorporationLarge multiplier for programmable logic device
US9098332Jun 1, 2012Aug 4, 2015Altera CorporationSpecialized processing block with fixed- and floating-point structures
US9189200Mar 14, 2013Nov 17, 2015Altera CorporationMultiple-precision processing block in a programmable integrated circuit device
US9207909Mar 8, 2013Dec 8, 2015Altera CorporationPolynomial calculations optimized for programmable integrated circuit device structures
US9244483Aug 8, 2011Jan 26, 2016Pentomics, Inc.Excess-fours processing in direct digital synthesizer implementations
US9268529May 14, 2013Feb 23, 2016Pentomics, Inc.Efficient angle rotator configured for dynamic adjustment
US9348795Jul 3, 2013May 24, 2016Altera CorporationProgrammable device using fixed and configurable logic to implement floating-point rounding
US9379687Feb 28, 2014Jun 28, 2016Altera CorporationPipelined systolic finite impulse response filter
US9395953Jun 10, 2014Jul 19, 2016Altera CorporationLarge multiplier for programmable logic device
US9547327Jan 25, 2016Jan 17, 2017Alan N. Willson, Jr.Excess-fours processing in direct digital synthesizer implementations
US9600278Jul 15, 2013Mar 21, 2017Altera CorporationProgrammable device using fixed and configurable logic to implement recursive trees
US20030005264 *Dec 29, 2000Jan 2, 2003Wolfram DrescherDevice and method for control of the data stream
US20040034641 *Aug 13, 2002Feb 19, 2004Steven TsengMethod and system for decimating an indexed set of data elements
US20070247189 *Dec 5, 2006Oct 25, 2007MathstarField programmable semiconductor object array integrated circuit
US20080183790 *Nov 9, 2007Jul 31, 2008Willson Jr Alan NEfficient Angle Rotator Configured for Dynamic Adjustment
US20100030831 *Aug 4, 2008Feb 4, 2010L-3 Communications Integrated Systems, L.P.Multi-fpga tree-based fft processor
CN100539707CJan 12, 2006Sep 9, 2009安凯(广州)软件技术有限公司Data driven reconfigurable color image processing system
CN101630308BJul 16, 2008Apr 17, 2013财团法人交大思源基金会Design and addressing method for any point number quick Fourier transformer based on memory
Classifications
U.S. Classification708/622, 708/440
International ClassificationG06F17/14
Cooperative ClassificationG06F17/142
European ClassificationG06F17/14F2
Legal Events
DateCodeEventDescription
Mar 14, 1997ASAssignment
Owner name: XILINX, INC., CALIFORNIA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:MINTZER, LESTER;REEL/FRAME:008446/0654
Effective date: 19970313
Apr 15, 2003FPAYFee payment
Year of fee payment: 4
May 17, 2007FPAYFee payment
Year of fee payment: 8
May 23, 2011FPAYFee payment
Year of fee payment: 12